Dear all,
I have data set for prevalence of mental illness for two years, from 2017-19 (36 months) and 2020-21 (14 months). I am comparing the prevalence of mental illness rate between 36 months and 14 months. For this I am comparing the mean mental illness rate between 36 months and 14 months through t-test using SPSS software and well as using the formula.
However, I have observed that the calculated mean value by using the formula given in statistics text book and the out put value given by SPSS software is varying considerably (0.47 vs 17). Now, I am confused which mean value should I give in the table, whether obtained through SPSS software or derived manually. Which one will be more reliable to give in the table? Please suggests.
Thanking you in advance.
Dr Bisu Singh
To clarify, the value given in the output of the t test is the mean difference, not the means of the groups. The means of each group will be included in the output under the label 'Group Statistics'.
The values should be approximately the same using either method. If they differ that much, a mistake was made either in the hand calculations or in SPSS. Without seeing the raw data, there's no way of knowing.
Hello Bisu,
To elaborate a bit on what Blaine Tomkins has already posted, spss does report group/batch means (for independent t-test), or measure means (for dependent t-test) in addition to the mean difference; so I'm a bit unclear as to what the problem is.
I can say that the basic algorithm for the t-test in spss hasn't changed appreciably in nearly 60 years, so I doubt that any discrepancy is due to a miscalculation in spss.
What could be going on is: you're using grouped data (rather than scores for individual cases), and entering values as if N = 1 (or each group has equal N, when in fact they don't).
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS